MO-CUPs present diagnostic and therapeutic challenges and have a poor prognosis with current treatments. This study aims to profile invasive mucinous tumours with and without ovarian involvement to better characterise MO-CUPs, differentiate between primary site and identify potentially actionable targets.
We analysed a de-identified, multi-platform tumour profiling dataset of mucinous tumours from CARIS Life Sciences. De-identified diagnostic pathology extracts were reviewed to assign tumour histology and extent of disease. We analysed results of a 45-gene next-generation sequencing (NGS) panel, gene amplification of HER2 and cMET and 17 immunohistochemistry (IHC) markers.
Invasive mucinous tumours from a cohort of 290 patients were eligible for analysis. A full dataset was available for 128, including 82 mucinous cancers of the ovary, 18 with extra-ovarian spread, and 28 of apparent non-ovarian origin (e.g. appendiceal). The most common mutations were in KRAS (63%), TP53 (35%), PIK3CA (13%) and PTEN (8%). The most common IHC findings included loss of expression in RRM1 (82%), ERCC1 (81%), TUBB3 (68%), thymidylate synthase (TS) (65%), and over-expression of EGFR (70%), TOP2A (66%), cMET (63%) and PGP (59%). Amplification of HER2 was seen in 9%, and cMET 1%.
In invasive cancers involving the ovary, PTEN and TP53 mutations were mutually exclusive (p=0.0002), and KRAS mutant cases differed with respect to lower expression of TS (p=0.01) and higher expression of TUBB3 (p=0.002). We found no statistically significant differences between mucinous cancers with and without ovarian involvement.
To our knowledge, this is the largest series of mucinous tumours involving the ovary reported to date. There are several potentially actionable biomarkers in this heterogeneous group of patients. The similarity in molecular profiling of “ovarian” mucinous cancers and non-ovarian mucinous cancers suggests that there is significant overlap and it may be more important to identify potential treatment targets agnostic of primary site.